A Lysa Phase II Study of Oral JAK1/2 Inhibitor Ruxolitinib in Advanced Relapsed/Refractory (R/R) Hodgkin Lymphoma (HL)

Background: JAK2 constitutive activation/overexpression is frequent in classical HL tumor cells and many autocrine/paracrine cytokines stimulate HL cells by recognizing JAK1- or JAK2-bound receptors. Thus, JAKs blockade may be of therapeutic value in HL.

Methods: A phase II study (HIJAK study)was conducted to evaluate safety and efficacy of ruxolitinib, an oral JAK1/2 inhibitor, in R/R HL, given at 20 mg bid for 6 cycles of 28 days. Dosage at 15 mg bid was planned for patients with platelets comprised between 75 and 200 x 10⁹/L at inclusion. Patients with platelets < 75 x 10⁹/L, neutrophils < 1000 x 10⁹/L were excluded. Maintenance beyond 6 cycles was permitted if disease control. The primary objective was overall response rate (ORR, Cheson 2007) at 6 cycles. Secondary objectives were safety, B symptoms relief, best ORR, response duration, PFS and OS. To be evaluable for response and survivals, patients had to receive at least one cycle of the study drug. The safety set included all patients who received at least one dose of ruxolitinib. Median follow-up was 27.1 months (95% CI: 14.4-27.1). Results: 33 patients were included between Jul 2013 and Dec 2014 in 10 LYSA centers in France and Belgium: M/F, 21/12; median age 37 (range 19-80) years; stage III/IV, 75.8%; B symptoms, 51.5%; median number of prior lines, 5 (range 1-16); prior transplantation, 60.6%; prior radiotherapy, 54,5%; prior brentuximab vedotin (BV), 82%; refractory to last therapy, 81.8%. Overall, median number of ruxolitinib cycles was 4. Nine (27.3%) patients received at least 6 cycles and 6 (18.2%) maintenance. ORR at the end of induction was 3/32 (9.4%) patients, all PRs. Best ORR at any time during study was 18.8% (6/32) with five PRs and 1 patient who converted into CR beyond 6 cycles. Transient stable disease was noted in 11 patients. Rapid and durable alleviation of B-symptoms (pruritus, fever, sweating) was frequently noted, especially pruritus which was present in 35.5% of patients before treatment and 6.6% of them after one cycle of ruxolitinib. Median duration of response was 7.7 months (95% CI: 1.8-NA). Two patients remain on therapy. Median PFS was 3.5 months (95%CI: 1.9-4.6) and median OS was 27.1 months (95%CI: 14.4-27.1).

Using bead-based immunoassays, plasma levels of 27 cytokines related to the immune system were measured at baseline and after cycle 1. Before ruxolitinib, there was no difference in cytokine levels between responders and non-responders. In responders, the only cytokine that significantly decreased was CX-CL10 (P.01). In patients presenting with pruritus (n=11), PDGF-BB, IL-5, IL-10, IL-12, IL-13, IL-17, eotaxin, FGF basic, MIP1b, rantes, and VEGF were significantly increased. In the latter patients, ruxolitinib treatment significanlty decreased PDGF-BB, IL-10, IL-12, IL-13, IL-17, FGF basic and VEGF. Among patients who were analyzable for JAK2 amplification in RS cells (n=12), polysomy was detected in all of them and specific JAK2 amplification in one. Further analysis of Jak2 targets by IHC will be performed.

40 adverse events (AEs) were reported in 14/33 patients (42.4%), of which 18 were related to ruxolitinib and 18 were grade ≥ 3. One AE led to permanent treatment discontinuation. No AE leading to death was reported. 87.5% of AEs recovered without sequelae. Eight SAEs (infection, 3; anemia, 1; diarrhea, 1; subdural hematoma, 1; bone pain, 1; pulmonary embolism, 1) were reported in 4 patients (12.1%), of which 2 were related to ruxolitinib. No grade 4 neutropenia and 1 grade 3/4 thrombocytopenia was observed. Five patients had grade 3 anemia. Twelve patients died due to lymphoma (83.3%) or toxicity of additional treatment (8.3%) or other reason (8.3%).

Among 30 patients who progressed (initial site in 97% and/or new site in 60%), 25 (83.3%) were retreated: with chemotherapy in 19 (comprising bendamustine in 10) and/or immunotherapy in 9 (rituximab, n=4; BV, n=3; nivolumab, n=2). Transplantation was eventually performed in 5/25 (4 allogenic, 1 autologous). In the 25 patients who were retreated, CR/PR rates were 10/15%, respectively. Conclusions: Ruxolitinib shows hints of activity beyond simple anti-inflammatory action in highly advanced, mostly refractory, HL patients, although most responses are short-lived. Toxicity was limited suggesting potential to be combined with other modalities. Further treatment, beyond ruxolitinib, was possible in most patients, even with chemotherapy.